Scholar
Tom Marty
Google Scholar ID: -YXor_wAAAAJ
Mila : Montreal Institute of Learning Algorithms
Citations & Impact
Citations
210
H-index
5
i10-index
4
Publications
12
Co-authors
14
list available
Publications
Resume (English only)
Academic Achievements
- Paper 'In-Context Learning and Occam's razor' accepted at ICML 2025 (April 2025).
- Paper 'WorkArena: How Capable are Web Agents at Solving Common Knowledge Work Tasks?' accepted at ICML 2024 (May 2024).
- First-author paper 'The Unsolved Challenges of LLMs as Generalist Web Agents: A Case Study' accepted at NeurIPS 2023 FMDM Workshop (November 2023).
- First-author paper 'Learning a Generic Value-Selection Heuristic Inside a Constraint Programming Solver' accepted at CP2023 (an A-grade conference) and awarded Distinguished Paper Award (July–August 2023).
- Journal paper 'Learning and fine-tuning a generic value-selection heuristic inside a constraint programming solver' accepted by Constraint Journal (November 2024).
- Awarded FRQNT doctoral scholarship for four years (March 2025).
- Received Mitacs Accelerate scholarship (2 units, $30K CAD) in March 2023 for visiting researcher work at ServiceNow.
Background
- Ph.D. student in Deep Learning at Mila under the supervision of Dhanya Sridhar.
- Originally from Réunion Island.
- Current research interests lie at the intersection of probabilistic modeling and learning theory, focusing on the role of simplicity and compression in designing robust learning algorithms.
- Specific topics include in-distribution and out-of-distribution generalization, amortized learning, meta-learning, generative modeling, and flow-matching—particularly for applications in cancer immunotherapy.
- Previously specialized in Reinforcement Learning and Combinatorial Optimization.
